2nd Place and 2nd Place Solution to Kaggle Landmark Recognition andRetrieval Competition 2019

TL;DR

This paper describes a comprehensive retrieval and recognition system for landmark identification, achieving second place in both Kaggle competitions through a multi-step process involving feature extraction, database augmentation, reranking, and voting.

Contribution

The paper introduces a complete landmark recognition and retrieval system utilizing PaddlePaddle, with novel integration of multiple techniques to improve accuracy and ranking performance.

Findings

Achieved 2nd place in Landmark Recognition 2019 Kaggle competition.

Achieved 2nd place in Landmark Retrieval 2019 Kaggle competition.

Demonstrated effectiveness of combined retrieval and recognition methods.

Abstract

We present a retrieval based system for landmark retrieval and recognition challenge.There are five parts in retrieval competition system, including feature extraction and matching to get candidates queue; database augmentation and query extension searching; reranking from recognition results and local feature matching. In recognition challenge including: landmark and non-landmark recognition, multiple recognition results voting and reranking using combination of recognition and retrieval results. All of models trained and predicted by PaddlePaddle framework. Using our method, we achieved 2nd place in the Google Landmark Recognition 2019 and 2nd place in the Google Landmark Retrieval 2019 on kaggle. The source code is available at here.